Technical writing on things I've learned building systems at scale: Postgres internals, .NET performance, data engineering patterns, and the occasional hard-won lesson.

Optimizing Levenshstein for Fuzzy Name Matching

Myers bit-parallel gave a 10x speedup over naive Levenshtein. Here's how a handful of cheap pruning tricks pushed it another 5x.

algorithmsperformancecsharpdotnetMay 2, 2026 · 11 min read

How Postgres CTIDs Gave Me a 30x Speedup on Processing 200 Million Rows

Indexes don't solve everything at scale. A story of how I cut 3-days worth of execution time down to a few hours by exploiting PostgreSQL's physical row addresses for cache-friendly sequential reads.

postgresqlperformancebig-datadotnetMarch 28, 2026 · 12 min read